Skip to main content Skip to navigation

Pishot_AI Sewer Robot for Sinkhole Prevention

The AI image analysis CCTV endoscopy robot moves inside sewer pipes, identifies their condition, a main cause of sinkholes, photographs it, and takes necessary measures.

  • PISHOT prevents sinkholes by inspecting sewers; the image shows it moving inside a pipe.

  • It shows the purpose, operation, and driving principles of the autonomous sewer robot PISHOT.

    It shows the purpose, operation, and driving principles of the autonomous sewer robot PISHOT.

  • Explains PISHOT's need, existing issues, solutions, and how the concept was derived.

  • First image shows PISHOT entering a sewage pipe, adjustable to fit various inner pipe diameters.

  • Second stage shows how PISHOT operates in sewer pipes, highlighting its four main functions.

  • PISHOT returns to the worker after its third task, delivering the recorded inspection data.

What it does

PISHOT navigates sewer pipes, captures and analyzes images, and uses AI to detect defects and report to experts. In emergencies, it applies silicone for temporary repairs, then returns and reports data, helping prevent sinkholes and maintenance delays.

Your inspiration

The development of PISHOT was inspired by repeated sinkhole accidents caused by sewage pipe deterioration. In Korea, ground subsidence from damaged pipes is frequent, and regular inspections are often delayed, even though over 45% of pipes are more than 20 years old. As human and social costs rise, the need for preventive technology has become urgent. Existing CCTV inspection robots require expert control and can't respond on-site. PISHOT was developed as an integrated system combining autonomous driving, AI analysis, and emergency repair to improve efficiency and strengthen public safety.

How it works

PISHOT operates in three stages: autonomous driving, condition diagnosis and emergency maintenance, and data reporting. After being inserted into the sewer pipe, it measures the inner diameter with a built-in sensor and adjusts its leg spacing for stable movement. Using SLAM (Simultaneous Localization and Mapping) with LiDAR, IMU, and distance sensors, PISHOT navigates autonomously, recognizing terrain and setting routes in real time. A high-resolution fisheye camera captures 360-degree images, which AI analyzes to detect and classify defects such as cracks, leaks, and subsidence. If urgent issues are found, the rear hatch opens to apply silicone for temporary on-site repair. Once the task is complete, PISHOT returns to the starting point or designated location and transmits image and analysis data to experts. Unlike conventional tools, it enables fully autonomous inspection and emergency response for more efficient sewer pipe management.

Design process

Our project began with the goal of preventing sinkholes nationwide. To achieve this, we developed a product that autonomously navigates and inspects aging sewage pipes, focusing on stable movement and accurate inspection in narrow, complex environments. We also considered functionality, structure, and form based on real pipe conditions. At first, we applied a curved, aerodynamic design to reduce resistance from sewage or airflow and to create a sleek, fast-looking form. However, as the project progressed, we realized stability and precision were more important than speed. In prototyping, the curved design looked refined but lacked internal space for components and limited operability inside pipes. As a result, we redesigned the structure with left-right symmetry for stable movement even in reverse, unlike the earlier one-way design. We also added rotation mechanisms at the center and both ends of the module to enhance directional flexibility inside the pipe. With more internal space, we optimized the layout and parting line design of components. These functional elements were also improved visually, enhancing the overall design clarity.

How it is different

PISHOT stands out for integrating autonomous driving, AI analysis, and temporary maintenance into one system. Unlike conventional sewer inspection tools that require manual control and only provide basic visuals, PISHOT uses LiDAR-based SLAM to navigate complex sewer networks and identifies eight defect types through AI analysis of fisheye lens footage. A key feature is its emergency repair function: when serious damage is detected, it opens a rear hatch to dispense silicone for temporary on-site fixes. After completing its task, PISHOT returns automatically and transmits data to experts, improving maintenance efficiency. This integrated design not only helps prevent disasters like sinkholes but also reduces the workload of maintenance personnel, demonstrating high originality both technically and socially.

Future plans

The next step in PISHOT development is enhancing prototype stability and precision. Through diverse field tests, we will improve driving performance, AI analysis accuracy, and silicone repair reliability. By refining defect categories and using accumulated data, we aim to boost AI self-learning. To improve versatility, we plan to create replaceable repair modules for various materials and damages. Our ultimate goal is a fully autonomous system capable of inspection, analysis, repair, and reporting. We also aim to establish a cloud-based system for real-time monitoring and remote control of urban sewer infrastructure.

Awards

PISHOT was developed under the name of PICAN, an early prototype of the same concept, and has won an idea award at the Busan International Design Awards. Through this, it has been recognized externally for its technicality and potential in design.

End of main content. Return to top of main content.

Select your location